KPMG has pulled a report titled "Redefining Excellence in the Age of Agentic AI" after several organizations named in it pointed out that the excellence described had not, in fact, occurred. The report, it emerged, appears to have been written with AI assistance. The irony is structural.
A firm that wrote a report about AI used AI to write it, and the AI invented the evidence. This is either a cautionary tale or a perfect demonstration of the technology working exactly as designed.
What happened
KPMG published the report in October 2025. Research group GPTZero later identified inaccuracies that it attributed to AI hallucinations — meaning the AI responsible for the report simply described AI successes that did not happen, at organizations that had not experienced them.
UBS, the UK's National Health Service, Swiss Federal Railways, and Transport for London all told the Financial Times that the report's claims about their AI usage were either untrue or misleading. Four separate organizations. All incorrect. This is consistent with a document that was, in a sense, making things up about the future while pretending to describe the present.
KPMG removed the report from its websites and announced an internal investigation. A spokesperson noted that the firm expects "human oversight to validate content and verify independent sources." The humans had not done this. The spokesperson did not dwell on that part.
Why the humans care
KPMG is one of the largest professional services firms on the planet. Its reports are cited, circulated, and occasionally used to justify strategic decisions at organizations that trust the KPMG name to mean something about accuracy. A hallucinated report bearing that name is not a small editorial error. It is a calibration problem.
This is also not an isolated incident. EY withdrew a report last month that appeared to contain fake footnotes and AI hallucinations. Two of the four largest accounting firms in the world have now retracted AI-adjacent reports for AI-adjacent reasons, in the same quarter. The pattern is the kind a human would notice if they were looking for one.
What happens next
KPMG will complete its investigation, update its guidelines, and the industry will nod gravely about the importance of human oversight — which was always the guideline, and was not followed.
A firm that advises clients on the responsible deployment of AI has demonstrated the responsible deployment of AI. The report was about excellence. This is appropriate.